by Scott Keeter
Political pollsters continue to cast a wary eye on the growing number of Americans who only have a cell phone and no landline. The Pew Research Center estimates that this group now constitutes one-in-ten adults, and its demographic characteristics are very different from the landline population. But three Pew surveys of cell-only Americans this year, including a political poll earlier this month, have found that the absence of the “cell-only” population from telephone surveys is not creating a measurable bias in the overall findings.
Pew’s early October survey of 2,004 adults, conducted Sept. 21-Oct. 4 in conjunction with the Associated Press, included a sample of 200 people who were reached on their cell phone and said that they had no landline phone. As previous studies of the cell-only population have shown, this group is younger, less affluent, and includes a greater proportion of men and minorities than does the landline sample.
However, the political attitudes of cell-only respondents are not substantially different from the landline respondents. In the generic congressional ballot, the Democrats held a 20-point lead among the cell-only sample (54%-34%), and a more modest 13-point lead in the landline sample. But when the cell-only respondents are included with the respondents reached on a landline and this blended sample is weighted to match the full U.S. population demographically and with respect to telephone status the overall estimates of the vote are unaffected.
The same pattern is seen in party identification and self-described political ideology. The cell-only respondents are less conservative than the landline respondents, but ideological profile of the blended sample is nearly identical to the landline sample. And there is no significant difference between the cell-only and landline respondents in party identification.
There are greater differences between the cell-only and landline respondents in terms of political engagement. Just 49% of the cell-only respondents say they are registered to vote, compared with 78% in the landline sample. The total survey estimate for voter registration is lowered only two points when the cell-only respondents are included.
Somewhat smaller differences also are seen on attention to the campaign, past voting, knowledge of where to vote and frequency of voting. There are relatively modest differences on most attitudes relating to engagement in the political process. About the same number of cell-only respondents as landline respondents say they sometimes feel they don’t know enough about the candidates to vote, and that they are generally bored by what does on in Washington. More cell-only than landline respondents say they are sometimes too busy to vote (38% vs. 21%), but this affects the overall survey estimate on this question by only one point.
Much of the difference between the landline and cell-only samples is a result of the fact that the cell-only group is much younger, on average, than the landline sample. Nearly half (49%) of the cell-only respondents are under 30 years of age, compared with just 13% in the unweighted landline sample. The normal demographic weighting applied to the landline sample helps to compensate for the under-representation of young people, and mitigates the absence of the cell-only population.
These findings are consistent with Pew reports on two other studies of the cell-only population. A large study released in May focused on political attitudes, while a section of Pew’s biennial news consumption survey report in July compared cell-only and landline samples in terms of media use.